354 Chapter Interference of m (Problem 9.25), whereas just the opposite applies to Newton's rings. An optical shop, in the business of making lenses, will have a set of precision spherical test plates or gauges. A designer can specify the surface accuracy of a new lens in terms of the number and regularity of the Newton rings that will be seen with a particular test gauge. The use of test plates in the manufacture of high-quality lenses, however, is giving way to far more sophisticated techniques involving laser interferometers (Section 9.8.4). 94.2 Mirrored Interterometers There are a good number of amplitude-splitting inter- ferometers that utilize arrangements of mirrors and beam-splitters. By far the best known and historically the most important of these is the Michelson inter- ferometer. Its configuration is illustrated in Fig. 9.24. An extended source (e.g., diffusing ground-glass plate illuminated by a discharge lamp) emits a wave, part of which travels to the right. The beam-splitter at O divides the wave into two, one segment traveling to the right and one up into the background. The two waves are reflected by mirrors M, and M; and return to the beam-splitter. Part of the wave coming from Mz passes through the beam-splitter going downward and part of the wave coming from M, is deflected by the beam- splitter toward the detector. Thus the two waves are united, and interference can be expected. Notice that one beam passes through O three times whereas the other traverses it only once. Consequently, each beam will pass through equal thicknesses of glass ‘only when a compensator plate C is inserted in the arm OM,. The compensator is an exact duplicate of the ws wh9-4 Amplitude-Splitting Interferometers 355 Figure 9.25 A conceptual rearrangement of the Michelson interferometer. beam-splitter, with the exception of any possible silver- ing or thin film coating on the beam-splitter. It is posi- tioned at an angle of 45°, so that O and C are parallel toeach other. the compensator in place, any optical path difference arises from the actual path difference. In addition, because of the dispersion of the beam- splitter, the optical path is a function of A. Accordingly, for quantitative work, the interferometer without the compensator plate can be used only with a quasimonochromatic source. The inclusion of a com- pensator negates the effect of dispersion, so that even fa source with a very broad bandwidth will generate discernible fringes. To understand how fringes are formed, refer to the construction shown in Fig. 9.25, where the physical O} components are represented more as mathematical sur- faces. An observer at the position of the detector will simultaneously see both mirrors My and M; along with the source 5 in the beam-splitter, Accordingly, we can redraw the interferometer as if all the elements were in a straight line. Here Mj corresponds to the image of mirror M; in the beam-splitter, and 3 has been swung over in line with O and M. The positions of these elements in the diagram depend on their relative di tances from O (e.g., Mj can be in front of, behind, or coincident with Mz and can even pass through it). The surfaces 2, and Ey are the images of the source 3 in mirrors M, and Mg, respectively. Now consider a single point S on the source emitting light in all directions; let's follow the course of one emerging ray. In actuality356 Chapter g Interference a wave from $ will be split at O, and its segments will thereafter be reflected by M, and Mg. In our schematic diagram we represent this by reflecting the ray off both Mz and M}. To an observer at D the two reflected rays will appear to have come from the image points 5, and Sg [note that all rays shown in (a) and (b) of Fig. 9.25 share a common plane of incidence]. For all practical purposes, S, and S2 are coherent point sources, and we can anticipate a flux-density distribution obeying Eq. (9.14). As the figure shows, the optical path difference for these rays is nearly 2d cos 0, which represents a phase difference of ko2d cos @. There is an additional phase term arising from the fact that the wave traversing the arm OMg is internally reflected in the beam-splitter, whereas the OM,-wave is externally reflected at O. If the beam-splitter is simply an uncoated glass plate, the relative phase shift resulting from the two reflections will (Section 4.5, p. 119) be a radians. Destructive, rather than constructive, interference will then exist when 2d COS Om = MAg, (9.44) where ™m is an integer. If this condition is fulfilled for the point S, then it will be equally well fulfilled for any point on = that lies on the circle of radius O'S, where O' is located on the axis of the detector. As illustrated in Fig. 9.26, an observer will see a circular fringe system concentric with the central axis of her eye's lens. Because of the small aperture of the eye, the observer will not be able to see the entire pattern without the use of a large lens near the beam-splitter to collect most of the emergent light.